一种基于混合映射策略的改进梅森旋转演算法  

A improved Mersenne Twister based on hybrid mapping strategy

作　　者：张琳琳 ZHANG Lin-lin(Public Security Information Technology Research Laboratory,The Third Research Institute of the Ministry of Public Security,Shanghai 201204,China)

机构地区：[1]公安部第三研究所治安信息技术研究室,上海201204

出　　处：《信息技术》2022年第10期12-17,23,共7页Information Technology

基　　金：国家重点研发计划项目(2017YFB0802300)。

摘　　要：文中基于Logistic与Kent两种混沌映射策略交叉映射初始序列的方法,提出一种改进的梅森旋转演算法(LKMT)。为考察其有效性,文中采用SP 800-22测试包测试包括频率、块频率、游程测试等随机性,并与文中提出的其他两种改进方法:基于Logistic混沌映射的MT方法(LMT)和基于Kent混沌映射的MT方法(KMT),进行分组实验对比分析,从不同类别的测试结果通过率比较,LKMT优于KMT和LMT,从通过测试的P_value数值大小来看,LKMT最优,但仅从近似商测试这一测试类别考察,KMT最优。综合考察,文中提出的LKMT相较于其他两种方法是较优的,且相对稳定的。In this paper, an improved Mersenne Twister(LKMT) is proposed based on the method of the Logistic and Kent chaotic mapping strategy’s cross mapping the initial sequence. To investigate its effectiveness, this paper adopts SP 800-22 test package to test the randomness, including the frequency, frequency within a block, RLins test etc, and compared with the other two improved methods, which are MT method based on Logistic chaotic mapping(LMT) and MT method based on Kent chaotic mapping(KMT). The result shows that LKMT is better than KMT and KMT is better than LMT in terms of the passing rate of different types of test results. In terms of P_value, LKMT is better than KMT and LMT, while only from the test category of approximate value, KMT is the best one. Comprehensive investigation shows that the LKMT proposed in this paper is better and relatively stable compared with the other two methods.

关 键 词：LOGISTIC混沌映射 Kent混沌映射 伪随机数 梅森旋转演算法 

分 类 号：TP30[自动化与计算机技术—计算机系统结构]